The invention relates to a method of forming high-k dielectric films such as hafnium or zirconium oxides or oxynitrides and their use for manufacturing semi-conductors.
With the shrink of the critical dimensions of the future generation of semi-conductor devices, the introduction of new materials, especially having high dielectric constant, is required. In CMOS architectures, high-k dielectrics are required to replace SiO2 which reaches its physical limits, having typically a SiO2 equivalent thickness of about 1 nm.
Similarly, high-k dielectrics are required in Metal-Insulator-Metal architectures for RAM applications. Various metal compositions have been considered to fulfill both the materials requirements (dielectric constant, leakage current, crystallization temperature, charge trapping) and the integration requirements (thermal stability at the interface, dry etching feasibility . . . ).
The Group IV based materials, such as HfO2, HfSiO4, ZrO2, ZrSiO4, HfZrO4, HfLnOx (Ln being selected from the group comprising scandium, yttrium and rare-earth elements) and more generally HfMOx and ZrMOx, M being an element selected from Group II, Group IIIa and Group IIIb, or a transition metal, are among most promising materials. Furthermore, Group IV metals composition can also be considered for electrode and/or Cu diffusion barrier applications, such as TiN for mid-gap metal gate and HfN, ZrN, HfSi, ZrSi, HfSiN, ZrSiN, TiSiN for MIM electrodes.
The main industrial options to enable the deposition of such thin films with a reasonable throughput and an acceptable purity are vapor phase deposition techniques, such as MOCVD (Metal-Organic Chemical Vapor Deposition) or ALD (Atomic Layer Deposition). Such deposition processes require metal precursors that must fulfill drastic requirements for a proper industrial use. Metal-organic or metal-halide precursors are required for those processes. Various hafnium and zirconium metal-organic compounds have been considered as precursors to enable such a deposition.
Halides such as HfCl4, ZrCl4 are the most common Hf/Zr precursors and have been widely described. Kim et al. disclosed the use of HfCl4 for the deposition of HfO2 by ALD (Kim et al., Electrochem Soc Proceedings 2005-05, 397, 2005). However, some by-products generated during the deposition process, such as HCI or Cl2, can cause surface/interface roughness that can be detrimental to the final properties. Other possible byproducts, depending on the oxygen source used, may be hazardous. For instance, OCl2, through the OCl fragment by QMS, has been detected as a byproduct of the reaction between HfCl4 and O3. Moreover, in the case of high-k oxide, Cl or F impurities are highly detrimental to the final electrical properties.
Triyoso et al. and Chang et al. studied the use of Hf(OtBu)4 for HfO2 MOCVD and ALD, respectively [Triyoso et al.; J. Electrochem. Soc., 152(3), G203-G209 (2005); Chang et al.; Electrochem. Solid. State Let., 7(6), F42-F44 (2004)]. Williams et al. have evaluated Hf(mmp)4 and Hf(OtBu)2(mmp)2 for MOCVD of HfO2. In WO2003035926, Jones et al. disclose solid Ti, Hf, Zr and La precursors improved with donor functionalized alkoxy ligand (1-methoxy-2-methyl-2-propanolate [OCMe2CH2OMe, mmp]) which helps inhibiting oligomerization of Zr and Hf alkoxide compounds and increasing their stability towards moisture. However, all those alkoxide precursors have the drawback not to enable self-limited deposition in ALD process as suggested by Potter et al. (R. J. Potter, P. R. Chalker, T. D. Manning, H. C. Aspinall, Y. F. Loo, A. C. Jones, L. M. Smith, G. W. Critchlow, M. Schumacher, Chem. Vap. Deposition, 2005, 11, N°3, 159-167).
Alkylamides precursors such as Hf(NEtMe)4, Hf(NMe2)4, Hf(NEt2)4 have been widely disclosed in the literature [Senzaki et al, J. Vac. Sci. Technol. A 22(4) July/August 2004; Haussmann et al, Chem. Mater. 2002, 14, 4350-4353; Kawahara et al., J. Appl. Phys., Vol 43, N°7A, 2004, pp 4129-4134; Hideaki et al., JP 2002-093804; Metzner et al. U.S. Pat. No. 6,858,547; Dip et al. US 2005/0056219 A1]. Group IV alkylam ides are both suitable for ALD and MOCVD processes. Furthermore, some are liquid at room temperature (Hf(NEt2)4 and Hf(NEtMe)4) and of sufficient volatility, and they allow self-limited ALD at low temperature for a limited thermal budget process. However, Group IV alkylamides, alkylamides in particular Zr compounds, have several drawbacks, among which they may decompose during the distribution to some extent leading to a possible clogging of the feeding line or the vaporizer, they may generate particles during deposition, they may entail non-uniform compositions during deep trenches deposition processes, and they only allow a narrow self-limited ALD temperature window, hence reducing the process window. In particular, Zr(NEtMe)4 may decompose in the distribution lines and generate particles above 170° C. which is a common distribution temperature. Hf(NEtMe)4 is more thermally stable yet do not afford self-limited atomic layer deposition above 300° C. due to thermal decomposition.
In WO 2007/055088, Thenappan et al. disclose hafnium and zirconium guanidinates complexes and their application for vapor phase deposition. Hf(NEt2)2[(NiPr-CNEt2]2 is given as example. Hafnium and zirconium guanidinates are however generally solids with a very limited volatility. As exemplified in thermal gravimetric analysis, one may not obtain Hf(NEt2)2[(NiPr-CNEt2]2 in vapour phase, without a risk of thermal decomposition and a subsequent particle generation.
Lehn et al. (Chem. Vap. Deposition, 2006, 12, 280-284) disclose tetrakis(trimethylhydrazido) zirconium [Zr(NMeNMe2)4,] and hafnium and their use for low temperature CVD. The exemplified compounds have an acceptable volatility (sublimation at 0.06 Torr, 90° C. reported) but they are solid at room temperature.
Carta et al. disclose the use of bis(cyclopentadienyl)bisdimethyl hafnium, [HfCp2Me2] (Carta et al. discloses in Electrochem Soc Proceedings, 260, 2005-09, 2005) and several authors (Codato et al., Chem Vapor Deposition, 159, 5 ,1995 ; Putkonen et al., J Mater Chem, 3141, 11, 2001 ; Niinisto et al., Langmuir, 7321, 21, 2005) proposed a new family of Zr and Hf compounds as alternatives to hafnium and zirconium alkylamides: Bis(cyclopentadienyl) bisdimethyl hafnium, bis(cyclopentadienyl) bisdimethyl zirconium, which allow an efficient ALD deposition process with an ALD window up to 400° C. and an achievement of films with less than 0.2% C in optimized conditions with H2O as co-reactant. However, HfCp2Me2 and ZrCp2Me2 both have the drawback of being solid products at room temperature (HfCp2Me2 melting point is 57.5° C.). This prevents IC makers to use those precursors in an industrial manner, that is using delocalized containers filling, and entail both facilitation and process issues.
In U.S. Pat. No. 6,743,473, Parkhe et al. disclose the use of (Cp(R)n)xMHy-x, to make a metal and/or a metal nitride layer, where M is selected from tantalum, vanadium, niobium and hafnium, Cp is cyclopentadienyl, R is an organic group. Only examples of tantalum and niobium cyclopentadienyl compounds are disclosed. However, no liquid precursor or a precursor having a melting point lower than 50° C. is disclosed.
Liquid bis(cyclopentadienyl) derivatives have recently been proposed by Heys et al. in WO 2006/131751 A1. However, they still present the disadvantage of limited volatility and also present large steric hindrance that may limit the achieved growth rate.
Today, there is a need for providing liquid or low melting point (<50° C.) group IV precursor compounds, and in particular Hf and Zr compounds, that would allow simultaneously a proper distribution (physical state, thermal stability at distribution temperatures), a wide self-limited ALD window, and a deposition of pure films either by ALD or MOCVD.